Import an Obsidian vault
Move a complete Obsidian folder into an internal CogLake workspace while preserving its useful structure and links.
Administrators and members with write access can import an Obsidian vault into an internal company or personal workspace. Open the plus menu beside file search, choose Import Obsidian vault, and select the vault folder.
CogLake creates a new conflict-free root folder and imports the vault without overwriting existing workspace content.
What is preserved
- nested folders and Markdown file names
- YAML frontmatter and normal Markdown links
- Obsidian wiki links such as
[[Runbook]],[[Runbook#Recovery]], and aliases - relative links between notes with the same file name in different folders
- inline image and video embeds such as
![[Media/overview.png|Overview]] - supported PDF, Office, image, audio, video, CSV, JSON, XML, and text attachments
Imported attachments follow the same API upload, malware scanning, processing, permission, and indexing pipeline as files added normally. They are never read directly from object storage by the browser.
Compatibility boundary
| Vault content | Import behavior |
|---|---|
| Markdown, frontmatter, wiki links, headings, and aliases | Preserved |
| Relative images and videos | Imported and rendered inline when permitted |
| Supported business files and media | Imported as workspace files |
.obsidian, themes, caches, and Git metadata | Ignored |
| Canvas files and unsupported plugin data | Ignored |
| Dataview queries and other plugin-generated output | Source Markdown remains, but dynamic plugin output is not reconstructed |
| Symbolic links | Not followed by the browser folder picker |
One import accepts up to 5,000 selected files, 100 MB of Markdown text, and 250 supported attachments. Larger vaults should be split into logical folders and imported in separate runs. If one attachment fails, imported notes and other attachments remain available and the final status names the failed files.
After import, internal Markdown notes remain editable and all successfully processed content becomes available to permission-aware search and agents.